Meta is trying to make Facebook suck less by simplifying things a bit

Australia has enacted a ban on social media access for individuals under the age of 16, effective December 10, 2025. This legislation targets major platforms such as Facebook, Instagram, and TikTok, marking one of the strictest measures globally to protect minors from online risks.

Google is set to launch its AI glasses next year, featuring audio and display options that promise screen-free assistance, live translations, and seamless smartphone integration. This move positions Google to compete directly with Meta's offerings in the augmented reality space.

A Border Patrol agent was recorded wearing Meta's Ray-Ban smart glasses during a raid, with the recording light clearly visible, despite a Department of Homeland Security (DHS) ban on officers using personal devices for recording. This incident raises concerns about compliance with regulations intended to protect privacy and civil liberties.

Meta has received approval from the European Union for its revised advertising experience, following penalties under the Digital Markets Act (DMA). This approval comes after the company faced scrutiny for its data practices and advertising models, which were deemed non-compliant with EU regulations.

A new dataset called EgoCampus has been introduced, focusing on egocentric pedestrian eye gaze during navigation in outdoor campus settings. This dataset includes recordings from over 80 pedestrians across 25 unique paths totaling 6 km, utilizing Meta's Project Aria glasses to capture eye tracking and environmental data.

The introduction of SPOT (Stopping Points in Online Threads) marks a significant advancement in the field of natural language processing, providing the first annotated French corpus designed to identify critical interventions in online discussions. This corpus consists of 43,305 annotated Facebook comments related to misinformation, offering a new lens through which to analyze online discourse.

A recent study has demonstrated the application of Post-Training Quantization (PTQ) to optimize Large Language Models (LLMs) for mobile execution, specifically focusing on Meta's Llama 3.2 3B model. The research achieved a 68.66% reduction in model size through 4-bit quantization, enabling efficient inference on Android devices using the Termux environment and the Ollama framework.

The study explores the trade-off between accuracy and computational efficiency in time series classification, highlighting the limitations of comprehensive ensembles like HIVE-COTE 2.0, which require extensive training time. By combining Hydra and Quant algorithms, the research evaluates performance across ten large-scale MONSTER datasets, achieving a mean accuracy improvement from 0.829 to 0.836 in seven datasets. However, the findings reveal a significant meta-learning optimization gap, with prediction-combination ensembles capturing only 11% of theoretical potential.